The emerging "information super-highway" and accelerating improvements in computer and mass storage technologies will soon make multimedia services, such as video-on-demand (VOD), a reality. Since a media server will be required to move a large amount of data over distribution networks, its I/O subsystem design is critical to its success. In this paper, we explore a new technology, Fibre Channel, as a storage interface for video servers. Specifically, we study Fibre Channel ioop topology that attaches multiple disk drives to a server host. In order to compare the performance of the new Fibre Channel interface with existing parallel SCSI, we built a video server simulator that simulates the behavior of both interfaces. The results show that with the same number of disks and the same system configurations, one Fibre Channel loop can achieve 50% higher performance (measured by the number of concurrent streams supported by the server) than four fastlwide SCSI channels. Our results show that Fibre Channel is attractive as a video/media server I/O interface. We also analyze the buffer size requirements and I/O transfer size for vanous configurations.
Abstract. In this article, we consider the problem of scheduling customers in a real-time system in which all customers are required to be serviced. Such a non-removal system can be distinguished from a removal real-time system in which customers can be removed prior to completing service. We describe and evaluate a simple paradigm for mapping policies for removal systems to policies for non-removal systems. We show that several policies known to be optimal for removal systems map into policies which are also optimal for non-removal systems. The article concludes with an application of this paradigm to scheduling requests with real-time constraints on disk subsystems.
L IntroductionIn this article, we study the problem of scheduling customers with deadlines in a non-removal real-time system (i.e., a system in which all customers must complete their service). Rather than propose and evaluate policies specifically for this system, we study the relationship between a non-removal real-time system and one in which customers may be removed prior to completing their service. We shall refer to the latter as a removal system (i.e., a system in which customers that miss their deadlines may be removed). We present a simple paradigm for transforming scheduling policies for removal systems to scheduling policies for nonremoval systems. We show how this paradigm can be used to obtain optimal policies for several classes of non-removal systems (i.e., policies that minimize the fraction of customers that miss their deadlines). In addition, we apply the paradigm to the problem of scheduling I/O requests with real-time constraints to disks when requests must always be completed. In doing so, not only do we present high performance disk scheduling policies for non-removal systems but also new policies for removal type disk systems that provide better performance than those previously described in the literature.Previous work on non-removal real-time systems has been concerned either with reducing the customer lateness and tardiness, with minimizing the customer loss ratio in the case that deadlines are not known, or minimizing the customer loss ratio when deadlines are known in the absence of arrivals. Su and Sevcik (Su and Sevcik 1978) looked at the problem of scheduling customers with deadlines in a queue. They showed that the earliest
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.